🎬 Introduction to Observability and Prometheus Tutorial
Video Details
Channel: NullSafe Architect
URL: https://www.youtube.com/watch?v=sNk9NkgTOLs
Relevance: ⭐⭐⭐
Summary
Foundation tutorial on software observability: the three pillars (metrics, traces, logs), Prometheus for metrics collection, Grafana for visualisation, and the difference between monitoring (is the system up?) and observability (why is the system behaving this way?). Includes a Docker Compose setup.
PUMA Relevance
Provides the theoretical foundation for PUMA’s observability approach. While PUMA uses Arize Phoenix (specialised for LLMs) rather than Prometheus (general infrastructure), the three pillars framework applies: metrics (F1-macro per condition), traces (Arize Phoenix agent decision chains), logs (Arize Phoenix prompt→response pairs). Useful background for PUMA’s Section 3 methodology.